Recognition System Of Chinese-sensible Code And Related Encrypting Algorithms Based On Camera Image

As a free two-dimensional code system with independent intellectual propertyrights, the advantage of Chinese-Sensible Code lies in its superior characterprocessing capability. So Chinese-Sensible Code is being more and more widelyused in the daily practice, and the research of Chinese-Sensible Code is graduallybecoming a hot spot. Chinese-Sensible Code is used instead of ordinary ciphertextletter code in order to store information on a VAT invoice. Besides the digitalinformation, this technique could also allow the storage of the names of goods,models and types, as well as units. So it could help effectively avoid the arbitraryinvoicing, false invoicing and other illegal acts. Currently, the technique ofcollecting invoices information using ordinary scanners and restoringChinese-Sensible Code information using decoding algorithm has become moreperfect. Since2012, the technology has been gradually put into use, and therecognition rate has increased from85%in the beginning to the current98%.However, many problems such as the prices and noise of equipments and thespeed of image acquisition, severely limit both the application scope and the futureof Chinese-Sensible Code. Therefore, the equipment acquiring Chinese-SensibleCode image by taking pictures using digital cameras begin to emerge in the market.Although this solves the questions above to some extent, some missing functions ofthe device increase the difficulty in decoding Chinese-Sensible Code greatly, andthus bring higher requirement for our decoding algorithms. This requires us toimprove the algorithm of processing scanner image in large scale. Theseimprovement focus on several aspect such as image acquisition and preprocessing.For example, in dealing with the red seal filter and uneven brightness problemencountered during image acquiring, we adopt the method of image stitching technology and white balance algorithm, and get a good result, with almost no lost inthe recognition rate of the algorithm.For some problems existing in the Chinese-Sensible Code recognition algorithm,we made some adjustment, including the computing of the tile angle and distortioncorrection of the area of code on the invoice, and the selection of interpolationfunctions. We also come up with a method for improvement. Specifically, we use thethe random Hough transformation to compute the tilt angle of Chinese-SensibleCode; In terms of the selection of interpolation function, we found that BicubicInterpolation appears more suitable than Bilinear Interpolation for theChinese-Sensible Code we acquired. It not only improves the accuracy, but alsodidn’t bring apparent difference. Besides, we also get good result in separating codeand figure by using the second difference and integral method.In the year2013, people heard the news that some code of train tickets weredecoded. We also express concern about the security of Chinese-Sensible Code, andthe aspect of the security should be paid attention. We code the Chinese-SensibleCode by the way of embedding the digital watermark. Directly encrypting theencoded information can effectively avoid the successful decoding of conventionalrecognition software. When encoding information by the use of Arnold scramblingencryption algorithm, its iteration is periodic, so it is important for effectiveness ofencryption to choose appropriate times of iteration.